Feature Request: Error Handling Conditions



I'm using KNIME to execute long-running, intricate workflows. The problem is that sometimes, even though the workflow design is sound, a node fails due to some internal (bug, unexpected data, ...) or external error (database connection disrupted, ...) event. So a workflow that's supposed to run over night, breaks off and the processing time is lost.


I would like to suggest the following addition to KNIME's already impressive functionality: Error-handling conditions:

  • Each node gets an additional 'On Error' tab
  • In this tab, the user can specify what is to happen, should an unexpected error occur
  • Condition examples: Reset and Repeat; Repeat Only; Wait X min/sec/hr and Repeat; Ignore and Continue Workflow; Send Mail to (mail adress), etc.


I think this, or something similar, would be a very important addition to increase the robustness of workflow execution in production environments.


Thanks for a great product


Hi Gideon,

thanks for those interesting suggestions! We will add a "try-catch" type set of nodes with KNIME 2.8 (due this summer). But the idea of having nodes re-try execution is also an interesting one but I am not sure if we can make that for 2.8 as well...

Cheers, Michael

Hi Gideon,


I use switch nodes like Empty Table Switch to perform simple error handling in workflows that are started by cronjobs. That is not as convenient as "try-catch" routines, but it works - at least as a workaround until KNIME 2.8 is released.




Hi Frank,

I use the Empty Table Switch as well, but it won't help with an internal error condition - when the node fails.


@Michael: I'm very much looking forward to the try-catch nodes, thanks. And thanks for considering the re-try suggestion. KNIME really needs some way of gracefully recovering from an error condition.


Best Regards



\o/ Can't wait for summer...